スキル image-outpainting
📦

image-outpainting

コンテンツリビジョン r1 中リスク ⚙️ 外部コマンド🌐 ネットワークアクセス📁 ファイルシステムへのアクセス

Extend Image Canvases with RunComfy

Cropped images often lack the surrounding space needed for banners, catalogs, or vertical posts. This skill guides RunComfy model selection and canvas-extension requests.

対応: Claude Codex Code(CC)
Maintain this Skill? Claim it 📊 69 十分

自分のエージェントでインストール

このリクエストをエージェントにコピーしてください。正規の Skill ページとマニフェストが含まれています。

エージェントリクエスト
Review the Skillstore skill "image-outpainting" from https://skillstore.io/skills/prime-skills-image-outpainting.md and its manifest at https://skillstore.io/api/skills/prime-skills-image-outpainting/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.

エージェントは引き続き計画を提示し、セキュリティポリシーで必要な確認を求める必要があります。

エージェントが読めるリソース

AI エージェント、クローラー、スクリプトがページ全体ではなく整理されたコンテキストを必要とする場合は、これらのリンクを使ってください。

テストする

「image-outpainting」を使用しています。 Plan a 16:9 banner from a square portrait, keeping the subject unchanged.

期待される結果:

  • Suggested route: Nano Banana 2 Edit for prompt-driven side expansion.
  • Requested canvas: 16:9, with matching environment added on the left and right.
  • Preservation request: retain the original subject, pose, and central content; inspect the generated result for changes.

「image-outpainting」を使用しています。 The expanded photograph has mismatched lighting. I have a reference image showing the desired surroundings.

期待される結果:

  • Suggested route: GPT Image 2 Edit with the original and reference image.
  • Request matching lighting and surrounding style while retaining the original visible area.
  • For finer boundary control, consider a listed ComfyUI workflow in the RunComfy GUI.

セキュリティ監査

中リスク

All 61 static alerts are false positives involving Markdown formatting, documentation links, protected token storage, or image-editing prose. A separate shell-quoting flaw breaks the portrait example and makes the blanket shell-safety assurance unreliable. No evidence found of malicious prompt injection or unauthorized data exfiltration; the external CLI implementation was not available for verification.

1
スキャンされたファイル
194
解析済み行数
0
レビュー項目
0
誤検知を無視

確認済みのセキュリティ上の懸念 (1)

中
Unsafe Shell Quoting and Misleading Input-Safety Guidance
The portrait prompt contains "subject's" inside a single-quoted shell argument, breaking the example. The claim "No shell-injection surface" overlooks shell parsing before CLI execution. Reusing this pattern with unescaped user text can expose shell metacharacters.
The apostrophe visibly terminates the shell quote, and the blanket safety claim ignores that parsing stage. No active attacker-controlled execution is demonstrated.
監査者: codex
このレポートを共有・引用

バージョン付き評価レポート、中立的なバッジ、埋め込みカード、引用を共有できます。Skillstore は証拠を報告しますが、この Skill が安全かどうかは判断しません。

バージョン別レポートを開く
セキュリティ評価

レポートリンクをコピー

https://skillstore.io/skills/prime-skills-image-outpainting/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdownバッジ

[![Skillstore security assessment](https://skillstore.io/badges/skills/prime-skills-image-outpainting/security.svg)](https://skillstore.io/skills/prime-skills-image-outpainting?utm_source=security_passport_badge)

HTMLバッジ

<a href="https://skillstore.io/skills/prime-skills-image-outpainting?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/prime-skills-image-outpainting/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

埋め込みカード

<iframe src="https://skillstore.io/embed/skills/prime-skills-image-outpainting.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>
学術引用 (APA · BibTeX · CFF)

APA形式の引用

prime-skills. (2026). image-outpainting security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/prime-skills-image-outpainting/audits/1

BibTeX形式の引用

@techreport{prime-skills-prime-skills-image-outpainting-2026, author = {prime-skills}, title = {image-outpainting security audit report (audit version 1)}, institution = {Skillstore}, year = {2026}, number = {1}, url = {https://skillstore.io/skills/prime-skills-image-outpainting/audits/1}, note = {Author version unspecified} }

CITATION.cff

cff-version: 1.2.0 message: "If you use this Skill, cite its author and this versioned security audit report." title: "image-outpainting security audit report (audit version 1)" version: "unspecified" type: report authors: - name: "prime-skills" date-released: "2026-11-04" url: "https://skillstore.io/skills/prime-skills-image-outpainting/audits/1" identifiers: - type: other value: "skillstore:prime-skills-image-outpainting:audit:1" description: "Skillstore immutable audit report identifier"

バリアントを比較

インストール可能なバリアント 4 件

各作者のスキルは個別のインストール項目として維持されます。推奨バリアントは Skillstore の証拠で順位付けされます。

このバリアントが首位の理由

Skillstore スコアが最高

agentspace-so-image-outpainting

Skillstore スコア 79
証拠の信頼度 高
Skillstore 利用状況 14
更新済み

2026-10-22

runcomfy-com-image-outpainting

Skillstore スコア 78
証拠の信頼度 高
Skillstore 利用状況 46
更新済み

2026-09-09

doany-ai-image-outpainting

Skillstore スコア 76
証拠の信頼度 高
Skillstore 利用状況 53
更新済み

2026-09-09

prime-skills-image-outpainting

Skillstore スコア 69
証拠の信頼度 中
Skillstore 利用状況 不明
更新済み

2026-11-05

Skillstore スコア

このスコアの理由 証拠の信頼度: 中
55
アーキテクチャ
85
保守性
87
コンテンツ
65
コミュニティ
83
仕様準拠

作成できるもの

Adapt Photos for Vertical Posts

Extend a square image above and below for a 9:16 composition while requesting preservation of the central subject.

Expand Product Photography

Add surrounding counter texture and background space while matching product lighting and shadow direction.

Match Reference Surroundings

Plan a reference-guided expansion of a cropped photograph with matching architectural details, lighting, or period style.

これらのプロンプトを試す

Expand a Square Photo
Extend [image URL] to 16:9 with matching background on both sides. Preserve the central subject and existing visible area.
Create a Vertical Composition
Expand [image URL] to 9:16 above and below. Match lighting and background. Set the aspect ratio explicitly and preserve the visible subject.
Use a Style Reference
Plan a GPT Image 2 Edit expansion of [original URL] using [reference URL] for surrounding style. Preserve original details and explain route selection.
Plan Staged Expansion
Plan two 30-50% expansion passes for [image URL] toward [target ratio]. Assess each result, use safe argument handling, and recommend alternatives for visible boundaries.

ベストプラクティス

  • Specify extension direction, target aspect ratio, matching visual details, and original-content preservation in each request.
  • Use only authorized source and reference images, and confirm permission before submitting them to an external service.
  • Serialize JSON safely and pass it without shell interpolation; review example quoting and inspect each generated result.

回避

  • Concatenating user prompts or URLs into shell command strings, or copying the broken portrait example unchanged.
  • Treating generated surroundings as recovered facts or assuming preservation prompts guarantee identical original pixels.
  • Expecting video processing, internal mask edits, or GUI-only workflows to run through the documented still-image CLI route.

よくある質問

What does image outpainting change?
It generates new surroundings beyond the existing image boundary. It does not recover the actual scene outside the original frame.
What must I install before using this skill?
You need the RunComfy CLI, authentication, and network access. The skill documents installation and login but does not include the CLI implementation.
Which edit route should I choose?
The guide defaults to Nano Banana 2 Edit, suggests GPT Image 2 Edit for references, and recommends FLUX Kontext Pro for preservation-focused requests.
Will my original image remain exactly unchanged?
That is not guaranteed. Use explicit preservation instructions and compare the generated image with the original.
Can this skill edit masked regions or videos?
Those tasks are outside its scope. It links to separate inpainting and video-outpainting skills.
Are image requests processed locally?
No. The documented CLI sends requests to RunComfy and downloads results. Submit only authorized images and correct shell quoting before running examples.

開発者情報

作成者

prime-skills

ライセンス

MIT

Skillstore リビジョン

r1

バージョンに関する注意

作者はバージョンを宣言していません。

参照

00df9b2cd39764d369af6d0aa6ba7a955b71e9c0

メンテナンスの新しさ

2026/11/5

利用状況

0 ダウンロード · 0 閲覧

ファイル構成

📄 SKILL.md

prime-skills のその他のスキル

すべて表示
すべて表示